This study compared the measurements of the Motor Assessment Scale (MAS) to that of the Fugl-Meyer Assessment (FMA), a reliable and valid test for motor function in stroke patients. Thirty-two patients (20 men, 12 women) with a mean age of 60 years, and a mean time since stroke of 64.5 days, were tested with the FMA and MAS on two consecutive days. The Spearman correlation coefficient for total FMA and total MAS scores was 0.96. For selected items, significant (p < 0.001) correlations ranged from 0.65 to 0.93, except for sitting balance (-0.10). Low negative correlations between sitting balance scores and other items (motor and sensation) were found only for the FMA test, suggesting that the FMA sitting balance test is not valid for measuring balance and is likely responsible for the low correlation. Comparison of scores (normalized in percent of maximal value) for corresponding items of the two instruments also indicated that the FMA measured a higher (Wilcoxon = p < 0.0001) level of motor recovery, (especially in more disabled patients), for both the upper (15.7%) and lower extremities (27.5%). Lastly, a cumulative frequency distribution analysis indicated that a larger proportion of patients was found in the lower class interval scores of the MAS in comparison to the FMA. These results (1) support the concurrent validity of the MAS for measuring motor recovery in acute stroke patients; (2) demonstrate the poor validity of the FMA sitting balance test, and (3) suggest that the FMA scale can better discriminate the level of motor recovery than the MAS in the early stage of recovery or in the more disabled subjects.